I thought it might be interesting to try this out on wikipedia3 http://labs.systemone.at/wikipedia3 RDF/XML which unzipped to just under 4GB. I am using the 30 day trial for saxon. So far all I get as a result is 'java.lang.OutOfMemoryError: Java heap space' even with the most simple function similar to the documentation. I have 4GB of RAM, 2 twin core processors and am using the ant saxon-xslt task in the latest build. The ANT_OPTS is set to -Xms268M -Xmx1280M -XX:MinHeapFreeRatio=20 -XX:+HeapDumpOnOutOfMemoryError and I am on windows XP. If anyone can inform me how to configure a machine to process this file I would be grateful? Is it better to use Linux for memory allocation? Is there a way I can give Java more memory to process this?